SambaNova Lightning Digest: Introducing DeepSeek R1 671B! TLDR: The big whale @DeepSeek R1 671B on Cloud Even more models, try @allen_ai
's Tülu 3 16K context length for Llama 3.3 Dev spotlights SambaNova speeds on @huggingface Upcoming events
COMPUTING
-
DeepSeek R1 671B and New LLM Models on SambaNova Cloud
By
–
-
DeepSeek R1 671B Large Language Model Now Available in Cloud
By
–
The Big Whale @deepseek_ai R1 671B Has Landed in the Cloud! It runs just shy of 200 tokens per second & is available today for anyone to try. For API access, sign up for early access here https://
bit.ly/3XbePmK -
DeepSeek-R1: Cost-Effective Enterprise AI Without Breaking Budget
By
–
📣 @Ava_AITECH shares our big news! @RodrigoLiang recently discussed why DeepSeek-R1 is a game-changer for enterprises seeking to implement cutting-edge #AI without breaking the budget.
— SambaNova (@SambaNovaAI) 21 février 2025
Check it out: https://t.co/hWiFlwQxPY https://t.co/b2aWsIxvGW@Ava_AITECH shares our big news! @RodrigoLiang recently discussed why DeepSeek-R1 is a game-changer for enterprises seeking to implement cutting-edge #AI without breaking the budget. Check it out: https://
computerweekly.com/news/366619398
/DeepSeek-R1-Budgeting-challenges-for-on-premise-deployments
… -
Topological Qubits: Quantum Computing Breakthrough Explained
By
–
Perfect use case for Think Deeper: what's a topological qubit? https://t.co/4z6Lur3H7c pic.twitter.com/MkwSGI7sAe
— Mustafa Suleyman (@mustafasuleyman) 21 février 2025Perfect use case for Think Deeper: what's a topological qubit? https://
aka.ms/AAu76rr -
SambaNova Achieves 198 Tokens Per Second With Custom AI Chips
By
–
Take it from @techradar SambaNova runs @deepseek_ai R1 at 198 tokens/sec using 16 custom chips! Our SN40L RDU chip is 3X faster & 5X more efficient than GPUs 5X speed boost coming soon, with 100X capacity on cloud. https://
bit.ly/3Xatj6k -
DeepSeek-R1 Achieves 95% Chip Reduction Industry-First Performance
By
–
@techradar reported that SambaNova’s @deepseek_ai
-R1 achieved industry-first #AI performance with 95% fewer chips. @RodrigoLiang told @Waynewill1
: “We’re collapsing inference costs and slashing hardware demands—from 40 racks to just one.” #GPU #DeepSeek -
Moonshine Support: Memory Alpha Improvements
By
–
4. Moonshine support – Improved Memory Alpha, which would allow ChatGPT to access past conversations (currently in a limited Alpha release)
-
Data Centers Embracing AI To Power Artificial Intelligence
By
–
How Data Centers Are Embracing AI To Power AI
#AI #AIio #BigData #ML #NLU #Futureofwork @HaroldSinnott @fogoros @iainljbrown @NandoDF @katecrawford @drhassanrashidi @YuHelenYu http://
ow.ly/kh4830sK3n6 -
Auto-manage context optimization with caching for efficient token usage
By
–
We’ve also adjusted how auto-manage context works. When caching is active, more of your chat history is included while keeping total points lower. This ensure you get more context for less. (3/5)
-
PyTorch Derived from Torch: Index Return Differences
By
–
i think you keep forgetting pytorch derived from torch, not numpy.
we fixed as much differences as we can, but torch always returned indices in torch